This research seeks to determine: (1) the prevalence of drug treatment; (2) the costs of drug treatment in a predictive model; (3) its impacts on the usage of other treatments in a behavioral model; (4) the policy implications for health resources planning. It will investigate the treated prevalence, cost, offset, demand, and choice of outpatient v. inpatient care for persons who are treated for drug disorders. The database permits analysis of drug treatments while taking account of the interrelated nature of the numerous ADM (alcoholism, drug, and mental disorder) treatments. Analysis of drug treatment which ignores the alcoholism and mental health comorbidities is seriously incomplete. The following topics will be studied: Treated Prevalence - The variations in insurance coverage will be examined to determine the extent to which health care coverage determines the diagnosis given. Substitution - The extent to which increased use of one treatment is associated with decreased (increased) usage in another, e.g., whether increased use of drug treatment associated with decreased use of non-ADM care. Treatment Location - Predictors of treatment location, e.g., outpatient v. inpatient v. provider's office. Costs - A generalized cost function including probability of treatment, location (inpatient v. outpatient), and costs conditional on probability and location. Demand - Substitution effects using treatment prices in a budget share analysis; e.g., whether different providers/settings are complements or substitutes. Episodes - Determining how characteristics of drug treatment episodes affect costs of care, e.g., how total costs of drug and non-drug-related care in a short intensive episode differ from the total costs in a longer, less intensive episode. Health Resources Planning - Implications of findings for providing guidance to care managers, e.g., imputing possibilities from a wide range of settings as a menu of possibilities available to all managers. A database of insurance claims from MEDSTAT Corporation will be analyzed. One of the most comprehensive sources of insurance claims data available, it covers over 2 million individuals from over 70 corporations, a wide range of coinsurance arrangements, and a wide variety of ADM and non-ADM treatments.